Eur. J. Biochem. 86, 361 -370 (1978)
Isolation and Chemical Characterization of the Enterobacterial Common Antigen
Daniela MANNEL and Hubert MAYER
Max-Planck-lnstitut fur Immunobiologie, Freiburg i. Br.
(Received November 14, 1977)
The isolation of the soluble form of the Enterobacteriaceae common antigen from Salmonella montevideo was achieved by a combination of the phenol/water method and the phenol/chloroform/
petroleum ether extraction procedure. A phenol-soluble fraction highly enriched in the antigen was obtained which was further purified by DEAE-cellulose chromatography. A fraction eluted with 0.9 M ammonium acetate/methanol was highly active in coating erythrocytes for passive hemaggluti- nation with antisera specific for enterobacterial common antigen and was an active inhibitor of the hemagglutinating system for this antigen.
Chemical analyses showed enterobacterial common antigen to be a linear polymer of 1,4-linked N-acetyl-D-glucosamine and N-acetyl-D-mannosaminuronic acid units esterified to a small extent by palmitic and acetic acids. These amino sugar and fatty acid components represent about 65 - 70
%
of the material. The presence of additional still unknown lipid components which might account for the missing 30% and for the low solubility of the isolated antigen in water is discussed. The molecular weight of the readily soluble part of the inhomogeneous isolated enterobacterial common antigen was determined by sedimentation studies in an analytical ultracentrifuge using methanol as solvent: it was found to be 2700.
Enterobacterial common antigen, sometimes ab- breviated as ECA [1], is an antigen shared by almost all wild-type strains of Enterobacteriaceae [2,3]. It was first detected and defined by Kunin et al. [4] by its reactivity with antisera against Escherichia coli 014 in passive hemagglutination. Kunin, and later other investigators, tried different methods for isola- tion of this antigen and its chemical characterization [5 - 101. A comparison of the reported data was recent- ly made, showing that its chemical identity is not yet established [l]. It is difficult to draw any conclusion as to its chemical nature from these data, since each group developed individual extraction procedures and employed different bacterial strains as a source for their material.
A property of this antigen which has also hampered its characterization is its occurrence in two different forms. One is the ubiquitously occurring haptenic
Abbreviations. ECA, enterobacterial common antigen; GlcN, D-glucosamine; ManN, D-mannosamine ; ManNUA, D-mannos- aminuronic acid; carbodiimide reagent, N-cyclohexyl-N-[2-(4-mor- pholiny1)-ethyl j-carbodiimide methyl p-toluene sulfonate.
Enzyme. Hexokinase or ATP : ~-hexose-6-phosphotransferase (EC 2.7.1.1).
or free form, while the other is the immunogenic form which is restricted to a few R mutants of defined genetic background. Early observations by Kunin, later extended and confirmed by Neter et al. [ l l ] and Suzuki et al. [6], indicated that in the immunogenic E. coli 014 enterobacterial common antigen is present in two forms one being soluble in 8 5 % aqueous ethanol and thus being separable from lipopolysaccha- ride and the other one insoluble in ethanol and not separable from 0-antigen [ 6 ] . More recent investiga- tions showed that enterobacterial common antigen is immunogenic in R mutants of E. coli and Shigella which have the complete R core structure of the coli R1 or R4 core types and possess a functional rjuL gene [12- 141. The dependence of this immunogenicity on a defined R-core structure and a functional trans- locase system is a strong indication for an enzymatic transfer of haptenic enterobacterial common antigen onto the R lipopolysaccharide thus rendering it im- munogenic.
Recently it was shown by Makela et al. [15] and Schmidt et al. [16] that defects in either of two ilv- linked genes (rfe and rfj result in the formation of ECA-negative mutants. An rfe defect also blocks
0-chain synthesis [17], therefore rfe- mutants are also R mutants, whereas r f f mutants are S and ECA- negative [l].
Since ECA- mutants were available from Salmo- nella montevideo which could be used as negative controls during its isolation, the S. montevideo wild- type strain SH94 was used as source of enterobacterial common antigen.
In this communication the isolation and chemical characterization of enterobacterial common antigen will be discussed. The antigenic and immunogenic properties of the isolated material will be described in the following paper. Its state in a number of defined mutants of Salmonella minnesota and S . typhimurium will be described elsewhere.
MATERIALS AND METHODS Chemicals
Crystalline a-benzyl-N-carbobenzoxy-D-mannos- aminuronic acid was kindly provided by Dr P. H.
Gross (Stockton, Ca.) and was used as source of the standard of ManNUA. As water-soluble carbodiimide N - cyclohexyl- N - [2- (4- morpholinyl) - ethyl] -carbodi- imide methyl p-toluene sulfonate was purchased from Fluka (Buchs, Switzerland). All other chemicals were commercially available and were of pA quality.
Bacterial Strains
S. montevideo wild-type strain SH94 was used for extraction of enterobacterial common antigen. It was obtained from the Central Public Health Laboratory, Helsinki, Finland. As source of enterobacterial com- mon antigen as indicator antigen, E . coli K-12 mutant D21e8 ( g u m mutant) [18] was always used. For pro- duction of antisera to enterobacterial common antigen the ECA-immunogenic strains E. coliO14: K7 (F2387), Shigella boydii type 3- (F3140) and E. coli 08- : K27- (F470) were used as with previous studies [19].
Bacteria were grown in a fermentor at 37 "C at a constant pH of 7.2. The detailed conditions for growth and the medium used have been described previously [20]. For labeling of the GlcN and mannosaminuronic acid residues of enterobacterial common antigen the method of Kiss [21] was adopted: 10ml of standard I medium (Merck, Darmstadt) containing 0.4% D-glucose was inoculated with 1 ml of an over- night culture of
s.
montevideo and with 50 pCi [1-3H]- GlcNAc. At the end of the logarithmic growth phase the cells were harvested by centrifugation (10 min at 800 x g ) and twice washed with saline.An tisera
Antisera were prepared as described previously [19]. Briefly, rabbits (New Zealand White) were intra-
venously immunized at 4-days intervals with increasing amounts (0.25,0.5 and 1.0 ml) of bacterial suspensions ( 1O1O cells/ml), which had previously been heated to 100 "C for 1 h and washed with saline.
Isolation and Purification
of Enterobacterial Common Antigen
Bacterial cells were extracted by a combination of the hot phenol/water and phenol/chloroform/
petroleum ether extraction methods. Phenol-killed bacteria were first extracted by 45
%
aqueous phenol at 68 "C as described by Westphal et al. [22]. The resulting aqueous phase, after dialysis and lyophilisa- tion, was further treated with a mixture of phenol/chloroform/petroleum ether (2/5/8) according to Ga- lanos et al. [23]. After removal of the volatile chloro- form and petroleum ether under reduced pressure, lipopolysaccharide was precipitated by adding some drops of water to the resulting phenol phase; entero- bacterial common antigen remains in solution and can be recovered from the phenol phase after extensive dialysis and lyophilisation. The resulting material, being highly enriched for this substance, was re- suspended in water and centrifuged at 105 000 x g for 4 h. The supernatant was taken for further purifica- tion by DEAE-cellulose chromatography. A column (20 x 2 cm) of DEAE-cellulose (DE-32, Whatman), previously equilibrated with 0.5 M ammonium acetate/
methanol buffer of pH 6.2 [24] was used. The material from the supernatant fraction was applied to the col- umn dissolved in 0.5 M ammonium acetate/methanol buffer and was stepwise eluted with a buffer of increas- ing concentration of ammonium acetate (0.5, 1 .O, and 1.5 M) in methanol. The resulting fractions were combined, dialyzed and/or electrodialyzed [25] and freeze-dried. Rechromatography of the middle frac- tion (eluate with 1 .O M ammonium acetate/methanol) with a buffer of 0.9 M ammonium acetate/methanol was monitored by differential refractometry (Lamidur Refractometer, Winopal Forschung, Hannover). The eluates were checked for enterobacterial common antigen content by Ouchterlony gel precipitation using a high-titered antiserum to this antigen.
Analytical Methods
Hydrolysis of enterobacterial common antigen was done with HCl using various conditions; for the non-reduced polymer a maximum release of sugar constituents was observed with 4 M HCI at 100 "C for 2 h. Amino sugars were separated and identified by high-voltage paper electrophoresis using a vertical chamber system at field strengths of 45 V/cm and the following buffers : (A) pyridine/formic acid/acetic acid/water ( 1 : 1.5 : 10 : 90, v/v) of pH 2.8, (B) pyridine/
D. Mannel and H. Mayer 363 acetic acid/water (10:4:86, v/v) of pH 5.3 and (C)
molybdate buffer of pH 5.0 [26].
Thin-layer chromatography was carried out on cellulose plates (DC cellulose on aluminium foil, Merck) with the following solvent : butan-I-ol/acetic acid/water (4: 1 : 5). Reducing sugars were identified by alkaline silver nitrate (Trevelyan reagent) and amino sugars also by ninhydrin. Lactones of uronic acids were detected by the hydroxamate reagent [27].
Analysis of neutral sugars was done by converting the sugars of the hydrolysate into the volatile alditol acetates and passing them over an ECNSS-M column.
Amino sugars and amino acids were separated and quantified on an amino acid analyzer (Beckman model 120B). Ester-linked and amide-linked fatty acids were characterized and quantified by gas-liquid chromatog- raphy according to Rietschel et al. [28]. Total 0-acetyl and N-acetyl groups were determined according to the method of Fromme and Beiharz [29].
Permethylation of the isolated fraction of entero- bacterial common antigen was done by the methyl- sulfinylcarbanion/methyl iodide reactant of Stellner et al. [30]. Gas-liquid chromatography/mass spectro- metry was performed on a Finnigan quadrupole model 3200 system coupled to a model 6000 data and graphic output system. A U-shaped glass column (0.2 x 152 cm) filled with ECNSS-M (3
%
on Gas- chrom Q) was used with a helium flow of 20 ml/min.The conditions were: 160 "C, 4"/min to a final temper- ature of 200 " C , electron energy 70 eV, mass range of mje = 40-400, integration time 7 ms per 1 mi..
3-Deoxy-~-manno-octulosonic acid was assayed by the thiobarbituric acid method [31], phosphorus ac- cording to Lowry et al. [32] and nitrogen by the Kjel- dahl test.
Calculation of the number of carboxylic groups in the polymer was obtained by titrating the electrodialyz- ed preparation of enterobacterial common antigen with 5 mM NaOH and recording the titration curve by an automatic titrator (Radiometer Copenhagen, Type TTT lb). The sedimentation coefficient was determined by the method of Yphantis [33] and the partial specific volume calculated according to Kratny et al. [34] using a Spinco analytical ultracentrifuge (model E with schlieren optics).
Reduction of Carboxyl Groups in the Polymer
Reduction of the carboxylic groups of the mannos- aminuronic acid residues in the polymer was achieved by the method of Taylor and Conrad [35]. Prior to reduction, the carboxylic groups were substituted by reaction with a water-soluble carbodiimide reagent, then the reduction was carried out with NaBH4.
Briefly, the uronic-acid-containing fraction was desalt- ed by treatment with Dowex 50/Hf and then dissolved in a small volume of water. The pH was brought to
4.75 and was maintained close to this value during the esterification by dropwise addition of 0.1 M HCl (Radiometer Copenhagen, type TTT 1 b). For carboxyl group substitution, 212 mg of the carbodiimide rea- gent were added as aqueous solution to 5 mg of the polymer-containing fraction. The mixture was allowed to react for 1 h at room temperature. For reduction, the substituted polymer was dissolved in water, a drop of octan-1-01 was added and then an aqueous solution of 2 M NaBH4 was added dropwise. The reaction mixture was kept at pH 7 for 1 h by auto- matic addition of 4 M HCl (Radiometer Copenhagen, type TTT lb). The reaction product was then dialyzed and lyophilized.
Alkali Treatment of Lipopolysaccharide
For sensitizing red blood cells, lipopolysaccharide was treated with alkali (0.25 M NaOH, 56 "C, 1 h) as described by Neter [36].
Serological Methods
Passive hemagglutination was performed with fresh human erythrocytes (blood group A), although glutardialdehyde-preserved human red blood cells gave comparable results. They were thoroughly wash- ed with saline and suspended to a concentration of 0.5
%
in saline or phosphate-buffered saline (NaCl/Pi) before addition of antigen. For sensitizing red blood cells with the determinant for enterobacterial common antigen, supernatants of heated cultures (1 h, 100 "C) or 50 pg of purified antigen were added to 5 ml of red blood cell suspension. Incubation was performed at 37 "C for 30 min, then the excess of antigen was washed off and the sensitized cells were resuspended in saline or NaCl/Pi to give a 0.5 sus- pension. The blood suspension was added to a series of antisera dilutions as described in detail [37]. For measuring the inhibitory capacity of a substance in the hemagglutination inhibition test the inhibiting substance was dissolved in normal rabbit serum (1 : 50 diluted with NaCl/Pi) and serial dilutions (25 pl) of the inhibitor substance ranging from 250-0.25 pg/
ml were prepared. 25 pl of the appropriate serum dilutions (2 - 3 hemagglutination units) were added.
After incubation at 37 "C for 1 h, 50 pl of the sensi- tized red cells were added and the plates again incu- bated for the same length of time. The lowest inhibitor concentration giving total inhibition of the hemag- glutination was measured after 1 h at room tempera- ture.
Agar Gel Precipitations
These were performed according to Ouchterlony [38] and immunoelectrophoresis in sodium barbi- turate buffer of pH 8.6 by the microtechnique of Scheidegger et al. [39].
Phosphorylation by A T P and Hexokinase
bacteria Smonfevideo S H ~ L Determination of the optical configurations of GlcN and ManN in hydrolysates of the carboxyl- reduced enterobacterial common antigen was per- formed with ATP and hexokinase as described recently
Phenol-killed 100% related to dry weight
phenol /water
_1
extract ion [401.RESULTS
water phase 10%
I
Isolation of Enterobacterial Common Antigenphenol/chloroform/
petroleum ether
water precipitate
A
phenol solublei
alkali treatmentdialysis
centrifugation centrifugation cent r i f ugat ion
L.9%
Fp%
0.3%fi
DEAE-
cellulose chromatography P k L- DE AE
PtL-DEAEalk
It has previously been observed that lipopoly- saccharide extracted from strains containing entero- bacterial common antigen by the phenol/water pro- cedure always contained the determinant of this antigen in addition to endotoxic lipopolysaccharide.
Lipopolysaccharide of R mutants containing entero- bacterial common antigen and isolated by the phenol/
chloroform/petroleum ether method was, however, found to be free of this antigen [12]. Attempts were therefore made to combine these two extraction pro- cedures for the isolation of enterobacterial common antigen. Similar techniques have also been used by Galanos et al. [41] for the purification of lipopoly- saccharide of enterobacterial S forms. With this meth- od (shown in Fig. 1) phenol-killed bacteria of Sal- monella montevideo SH 94 were used for the extraction.
Purified lipopolysaccharide, free of any detectable specificity for enterobacterial ~OmmOn antigen, was obtained in a yield of 5 % (of bacterial dry weight) in the GL-S fraction (Table 1). From the phenol phase a fraction PL-L was obtained as the supernatant fraction after extensive dialysis, lyophilisation and ultracentrifugation (105000 x g, 4 h). The material
Fig. 1. Scheme of extraction und purification procedure of entero- bacterial common antigen obtained f r o m S. montevideo SH94. Frac- tion GU-S = sediment and fraction GU-L = supernatant of the fraction insoluble in phenol/chloroform/petroleum ether; fraction GL-S = sediment and fraction GL-L = supernatant of the water precipitate; fraction PL-S = sediment and fraction PL-L = super- natant of the phenol-soluble fraction; all these fractions were ob- tained after centrifugation at 105000 x g for 4 h
Table 1. Distribution of serological 0 and enterobacterial-common-antigen activities in dfferent fractions obtained from S. montevideo SH94 (see Fig. I )
Titers in the passive hemagglutination test and hemagglutination inhibition test with antisera specific for anterobacterial common antigen (ECA) and 0-antigen are shown. ECA-antiserum = antiserum directed against Shigella boydii type 3- ; 0-antiserum = antiserum directed against S. montevideo SH94; ECA-antiserum 1 : 640 and ECA-containing lipopolysaccharide derived from E. coli K12 D21e8 were used for the inhibition test and the latter also as indicata antigen for the passive hemagglutination
Fraction Passive test Inhibition test
~~
~ ~~
~-
ECA-antiserum 0-antiserum ECA-antiserum
1 : 640
_ _ ~~
GL-L GL-S
5 120 2 560
< 10 < 10
~~ ~ _ _ _ _ _ _ _
~~
GU-L 5 120 2 560
GU-S 2 10240 5 120
< 10
< 10
1280 160
3.9 2 250
< 10
< 10
80 320
0.25 15.6 PL-L
PL-s
2 10240 2 10240 2560
2 10240 5 120 2560
40 40
< 0.25
< 0.25
D. Mannel and H. Mayer 365 amounted to about 0.3% of bacterial dry weight,
and was found by serological tests (hemagglutination inhibition) to be highly enriched for enterobacterial common antigen (Table 1). It did, however, still show a slight 0-titer when tested against the homologous S. montevideo SH 94 antiserum. Examination of the PL-L fraction in immunoelectrophoresis with a high- titered antiserum to enterobacterial common antigen revealed a precipitation line towards the anode in- dicating that the antigen was highly negatively charged.
Although a high inhibition of an hemagglutinating system for enterobacterial common antigen was also obtained with the sediment fraction after ultracentri- fugation (fraction PL-S in Fig. l), most of the further work was performed with the PL-L fraction.
Purification of Enterobacteriaf Common Antigen Ion-exchange chromatography on DEAE-cellulose was used as a further step in purification of the nega- tively charged antigen. With the knowledge that this antigen contains a high percentage of amino sugars, the PL-L fraction was extracted from bacteria labeled by growth in [l-3H]GlcNAc-containing medium [21].
The elution profile of the material with ammonium acetate/methanol buffer of increasing concentration (0.1 - 1.6 M) is shown in Fig. 2. The peak of the eluted 3H content coincided with the precipitation pattern obtained by agar double diffusion of the individual fractions. The bulk of the antigen elutes with 0.8- 0.9 M ammonium acetate in methanol. For prepara- tive purposes a stepwise elution with ammonium acetate of increasing concentration (0.5,l .O and 1.6 M) was carried out, yielding the entire enterobacterial common antigen material in the middle fraction.
The first fraction contained some lipophilic material without serological specificity for enterobacterial com- mon antigen and was not further examined.
Rechromatography of the middle fraction with 0.8 M ammonium acetate buffer eluted enterobac- terial common antigen as a symmetrical peak moni- tored by differential refractometry. This purified fraction (PL-L-DEAE) contained all the serological properties of the native antigen, i.e. erythrocyte- coating capacity, hemagglutination with its antisera, inhibition of its hemagglutinating system, as well as agar gel precipitation with high-titered antisera to enterobacterial common antigen. This material was used for the analytical and structural studies described below.
Chemical Analyses
Qualitative chemical studies showed that entero- bacterial common antigen consists primarily of D- glucosamine (GlcN) and D-mannosaminuronic acid (ManNUA) both being N-acetylated. A direct quan-
0 10 20 30
F r a c t i o n number
Fig. 2. Column Chromatography on DEAE-rellulose of fraction PL-L containing enterobacterial common antigenfrom S. montevideo SH94 labelled with 3 H . The antigen was eluted from the column (2 x 20 cm, DE 32, Whatman) at room temperature with a linear gradient (-) ranging from 0.1 - 1.6 M ammonium acetate.
( x ~ x ) N-Acety~-[l-3H]g~ucosdmine incorporated
tification of the aminuronic acid was not possible due to its rapid degradation under the strong acidic conditions used for its liberation [42]. After hydrolysis with hydrochloric acid the majority of liberated amin- uronic acid is present as lactone [40] which is easily demonstrable by high-voltage paper electrophoresis, by staining with either Trevelyan reagent, ninhydrin or the hydroxamate reagent (Fig. 3). After N-acetyla- tion of the hydrolysate a negatively charged component in buffer A was detectable with W Z G ~ ~ N = 0.27. The migration data in buffer systems A, B and C were the same as described recently for mannosaminuronic acid. Buffer system C readily separates glucosamin- uronic, galactosaminuronic and mannosaminuronic acids [26]. Final proof of the chemical identity of the aminuronic acid of enterobacterial common antigen with mannosaminuronic acid was obtained by co- electrophoresis with an authentic standard of D-
mannosaminuronic acid and by reducing the amin- uronic acid in the polymer to D-mannosamine. For quantitative determination of mannosaminuronic acid the polymer was reduced with NaBH4 after reaction with the carbodiimide reagent (repeated four times) and hydrolysis of the reduced polymer with 4 M HC1 for various periods of time. Fig. 4 shows that a maximal release of ManN was obtained after 6 h, and that GlcN and ManN are present in approximately a 1 : 1 ratio.
Both sugars in the reduced polymer were found to be substrates for enzymatic phosphorylation by hexokinase/ATP [43] showing the D-configuration of
ManNUA l a c t o n e
-
Glc N
+
ManNUA
+
S t a r t 4
PL-L GIC G l c N
Fig. 3 . High-voltage electropherogram of fraction PL-L from the wild-type S . montevideo SH94. 50 pg PL-L material was hydrolyzed with 4 M HCI at 100 "C for 2 h. After electrophoresis (3 kV) in buffer system (A) at pH 2.8 the electropherogram was stained with alkaline silver nitrate
GlcN and ManN and also of mannosaminuronic acid in enterobacterial common antigen.
The amount of amino sugar found by these analyses was 33% of the dried PL-L-DEAE fraction. This value is surely the minimal content because neither the reduction of ManNUA to ManN nor the libera- tion of the amino sugars by acid hydrolysis can be complete.
Investigation of the PL-L-DEAE fraction for neu- tral sugars (by gas-liquid chromatography) or for 3-deoxy-~-manno-octu~osonic acid (thiobarbituric acid reaction) was completely negative. Amino acids were found only in trace amounts, no amino acid being found preferentially. Fatty acids were present in a rather small percentage (Table 2) ; the main (ester- linked) fatty acid was palmitic acid.
Total nitrogen was determined as 3.58
%.
Assuming that the amino sugars are the only nitrogen-containing constituents of enterobacterial common antigen and that they occurred in a 1 : 1 ratio, the total percentage of amino sugars can be calculated as 48.8 "/,. This value is the upper limit of the amino sugar content and it agrees with a calculation based on the N-acetyl deter- mination. Assuming all amino sugars are N-acetylated, the N-acetyl content of 11.7%
is then equivalent to 45.6% of the amino sugar mixture.' ManN
' 0
0
0 1 2 4 6 8 10 12 16 20
Time (h)
Fig.4. Kinetics of the liberation of' GlcN and ManN f r o m the carhoxyl-reduced enterobacterial common antigen from S . monte- video SH94 by 4 M H C l at 100 "C. Reduction was performed ac- cording to the method described by Taylor and Conrad [35]
Table 2. Chemical composition of enterobacterial common antigen derivedfrom S . montevideo SH94
Amounts are given as percentages of the material dry weight.
n.d. = not detectable
Constituent Amount Remarks
7"
Amino sugars 45.6-50 D-G~cN, o-ManNUA (in a 1 : 1 ratio) Acetyl groups 13 11.7 N-acetyl,
1.3 % 0-acetyl
mostly C16:" ester-bound (2.0%)
Total fatty acids 2.5
Phosphate 0.47
Neutral sugars n.d.
Amino acids n.d.
Titration of the carboxylic groups of mannos- aminuronic acid in the PL-L-DEAE fraction (after electrodialysis) gave a value corresponding to 23 - 27
%
mannosaminuronic acid (three independent mea- surements). This is approximately half the amount of total amino sugar as calculated by nitrogen or N - acetyl determination (see above), a value which sup- ports the sugar analysis of the reduced polymer which gave an approximate ratio of 1 : l for GlcN and ManNUA.Analysis of the Amino Sugar Linkages
To determine the linkages of the two amino sugar constituents, a methylation analysis according to Hakomori was carried out with the carboxyl-reduced polymer. The partially methylated amino sugars in the hydrolysate were converted to alditol acetates and separated on an ECNSS-M column. Two peaks,
D. Mannel and H. Mayer 367
A
Time (arbitrary units)
B
C H ~ O A C
- 20 I
Hy-Nci:3
1581116202 CH30-CH 233
HF-OAc 4-
_ _ _ -
f
m 50 100 150 200 2 50
m m / e
r c a
f
C
50
0
50 100 150 200 250
m / e
Fig. 5 . Separation of partially methylated alditol acetates according to the method of’ Stellner et al. (301 on gas-liquid chromatography ( E C N S S - M c o l u m n ) at 160 “C (program 4 “Cjmin to 200 “ C ) . (A) Mass fragmentography of mass mje 158; (B, C) mass spectra and fragmenta- tion schemes of 3,6-di-O-methyl-l,4,5-tri-O-acetyl-2-deoxy-2-N-methylacetamidoglucitol (B) and 3,6-di-O-methyl- l ,4,5-tri-O-acetyl-2-deoxy- 2-N-methylacetamido-mannitol (C). The mass spectra were taken at an ionizing energy of 70 eV with a Finnigan quadrupol model 3200 system
eluted at 33.3 and 40.4 min, were found by monitoring the characteristic mass m/e 158. The spectra of the two components were almost the same showing that
both were derived from 3,6-di-O-methyl-l,4,5-tri-O- acetyl-2-deoxy-2-N-acetamido-hexitols (Fig. 5). NO
permethylated amino sugar and no mono-0-methyl derivative of amino sugars could be detected, indicat- ing that enterobacterial common antigen consists predominantly of a linear chain of 1,6linked N-acetyl- amino sugars (Fig.6)’. The lack of a permethylated (terminal) amino sugar is surprising considering the
date oxidation of the reduced polymer shows, however, that ManN is partly destroyed (27 ”/,), whereas GlcN remains unaffected. We were not able to render the reduced polymer completely periodate-oxidizable by a preceding hydrazinolysis. Periodate oxidation of the unreduced hydrazinolyzed polymer and a subsequent mild acid hydrolysis followed by NaBH4 reduction, afforded some erythronic acid, but no glyceric acid (Table 3) again pointing to a 1,4-linkage of the amin- uronic acid.
low molecular weight (about 2700) (see below). Perio-
In the nuclear-magnetic-resonance spectrum of alkali-treated antigen (after HjZH exchange) in ‘H20 a broad signal with an intensity of 1 is found at 6 = 5.33 ppm. The signals of equatorial anomeric protons are always in this region, hence one of the two glycosidic linkages expected for the antigen is ct and the other one must therefore be p. Further studies are needed for a more detailed
Isolation ofOligosaccharides
From enterobacterial common antigen subjected to 4 M HC1 hydrolysis at 100 oc, two oligo- saccharides could be separated by preparative high- voltage electrophoresis. These oligosaccharides Seem to be disaccharides. Both consist exclusively of GlcN
determination of the anomeric structures.
\
'0
-
c=o I H3
1
.
D - G I C N A C J A D - M ~ ~ N A C U A
- 6 Fig. 6. Proposed structure for the sugur moiety of enterobucteriul common antigen obtained from S. montevideo SH94
Table 3. Comparison of the electrophoretic and chromatographic behaviour of glyceric acid, erythronic acid and the fragment of enterobacterial common antigen obtained after periodate oxidation, mild hydrolysis and NuBH4 reduction
Migration relative to dOclA (3-deoxy-~-manno-octu~osonic acid), mdOclA, was measured by high-voltage electrophoresis (3 kV, 45 V/
cm) in system A at pH 2.8 and in system B at pH 5.3. RF values were measured by thin-layer chromatography on cellulose plates in butan-I-ol/acetic acid/water (4/1/5)
Substance mdOclA in system RF
~~ ~
A (pH 2.8) B (pH 5.3)
dOclA 1 1 -
Glyceric acid 1.07 1.47 0.32
Erythronic acid 0.89 1.37 0.4
Antigen fragment 0.89 1.36 0.4
and ManNUA as shown by further hydrolysis of the eluted fractions. One oligosaccharides (a) is readily liberated after short hydrolysis (1 h) with 4 M HC1 at 100 "C. It has a migration value of m G l c N = 1.35 in high-voltage electrophoresis at pH 5.3 and contains ManNUA in the lactone form as judged from the po- sitive hydroxamate test. Reduction of oligosaccharide (a) with NaBH4 and a subsequent hydrolysis and per- acetylation does not give glucosaminitol peracetate but glucosamine peracetate, proving that GlcN is not the reducing terminal of oligosaccharide (a).
Treatment of oligosaccharide (a) with 0.5 M NaOH (10 min, 25 "C) easily liberates GlcN and a reaction product showing a significant absorption at 300 nm.
The second oligosaccharide (b) can be isolated after strong acid hydrolysis (4 M HCl, 100 "C, 6 h) by high- voltage electrophoresis in buffer A with m G l c N being 0.72. It is very acid-stable and even after drastic con- ditions (6 M HC1, 100 "C, 32 h) it is not completely hydrolyzed. GlcN forms the reducing end as demon- strated by NaBH4 reduction. This oligosaccharide is therefore similar to aldobiuronic acids which are known for their extreme stability to acid hydrolysis.
Physico-chemical Proper ties of Enterobacterial Common Antigen
The isolated material was not soluble in water at concentrations higher than 0.3
%.
After removal of the ester-linked fatty acids by alkali treatment the solubility was only slightly increased to 0.5%. The material is, however, soluble in methanol. For this reason, the sedimentation studies on the untreated enterobacterial common antigen in the analytical ultracentrifuge were carried out in methanol as solvent.The sedimentation coefficient in methanol determined by the method of Yphantis [34] is s% = 1.5 S and that calculated in water is $0 = 0.56 S. The partial specific volume of the substance was measured [35]
as 0.5759 ml/g at 20 "C. From these data a molecular weight of 2700 was calculated for the part of the PL-
L-DEAE material which was readily soluble in meth- anol. It is obvious that enterobacterial common anti- gen in aqueous solutions forms micelles and co- micelles of much higher molecular weight.
DISCUSSION
The data presented in this paper show that the enterobacterial common antigen is a linear heteropoly- mer composed predominantly of alternating units of N- acetyl-D-glucosamine and N-acetyl-D-mannosamin- uronic acid which are partly esterified by palmitic acid. The high negative charge is due to the content of the aminuronic acid. At least 50% of the isolated and purified material consists of these two N-acetyl- amino sugars which occur in a molar ratio of about 1 : 1. The amino sugars are arranged in a chain with 1,4-linkages. Oligosaccharides, probably disacchari- des, consisting of both GlcN and ManNUA were isolated. Only one had GlcN as the reducing terminal, the other one had it as the non-reducing terminal.
This suggests that these two sugars occur in an alter- nating sequence in the polymer chain. As expected for 1,4-1inked N-acetyl-amino sugars, periodate oxi- dation had no effect. We have, however, no explana- tion so far as to why the amino sugars are not com- pletely destroyed by periodate oxidation after pre- ceding hydrazinolysis. Incomplete de-N-acetylation [44] or steric hindrance by the bulky carboxyl groups [45] might be responsible for the lack of reactivity.
The observed partial destruction of ManN ( z 27
%)
in the reduced polymer could indicate that either some of the ManNUA residues in the antigen chain are in a terminal position or carry unsubstituted -NH2 groups. The extent of the oxidation would be con- sistent with a structure having 8 - 10 amino sugars in a 1,4-linked chain, which is also indicated by the low molecular weight of 2700. On the other hand no permethylated mannosamine could be detect-
D. Mannel and H. Mayer 369 ed in the methylated reduced sugar polymer. The
material used for the methylation study was, however, only reduced once and it is possible that the mannos- amine obtained by a single reduction step gives so little permethylated material that it is below the limit of detection. It should be pointed out that the M , of 2700 is that of the readily soluble fraction of the inhomogeneous antigen isolated in methanol. In aqueous solutions it forms micelles of much higher molecular weight.
The partial esterification of enterobacterial com- mon antigen by palmitic acid is (partly) responsible for its hydrophobic character and may be necessary for its incorporation into the outer membrane of Enterobacteriaceae. Its low solubility in water, even after removal of the ester-linked fatty acids, suggests the presence of a hitherto undetected lipid component.
The addition of all components detected so far shows a deficit of about 30% on a weight basis. Marx et al.
[lo] described a cephaline-like L-phosphoglyceride as a constituent of their isolated material from S.
typhimurium. A careful search for cephaline-like sub- stances in our material was not successful.
Fig.6 shows the structure proposed for the sugar moiety of enterobacterial common antigen. No in- formation is available so far on the point of attachment of the palmitic acid. From the molecular weight and the percentage of palmitic acid present, it is obvious that not more than a single palmitic acid can be linked to one polymer molecule. Nevertheless splitting off this fatty acid leads to a complete loss of its erythro- cyte-coating abilities [12]. The structure proposed for the sugar part (Fig. 6) shows that p-elimination is theoretically possible and is expected. Indeed, it was found that the one disaccharide isolated after short hydrolysis (4 M HC1, 1 h) immediately liberated the terminal GlcN under alkaline conditions and gave a second reaction product with an absorption in the 300-nm range [46,47]. It is therefore assumed that
p-
elimination can also occur during alkali treatment, when some of the carboxylic groups of the ManNUA residues are esterified. This may be the reason for the lower inhibitory capacity of alkali-treated enterobac- terial common antigen observed in an hemagglutina- tion system (discussed in the following paper).
The chemical composition described in this paper does not agree with earlier reports on the chemical nature of enterobacterial common antigen. Most investigators found, however, that the Kunin antigen was negatively charged [5,8,10,12] and that it contain- ed considerable amounts of glucosamine [ S , 7,101. A sample of material isolated, extracted and purified according to the method of Marx et al. [lo] was kindly provided for us; we have demonstrated that it contained appreciable amounts of ManNUA and GlcN, indicating that part of the material may be identical to our PL-L-DEAE fraction. The high
hexose content described by Johns et al. [8] for their serologically uniform preparation, is not in accord with our isolated material, which was essentially free of neutral sugars. Since mutants with enzymatic defects in glucose, galactose and mannose biosynthesis (UDPG-pyrophosphorylase-less mutant Gal 23 K - , UDPGal-4-epimerase-less mutant PL-2 of E. coli K-12, and phosphomannose-isomerase-less mutant of S . typhimurium) all synthesize enterobacterial com- mon antigen in normal amounts and contain the GlcN/ManNUA polymer (data not shown) to an appreciable extent, these neutral sugars are therefore not essential parts of the antigen.
From these facts it might be assumed that not only a single antigen is shared by several Enterobacteria- ceae but the presence of at least two have to be con- sidered. In a following paper we will show that the serological and immunological properties of our isolated material are identical to that of the common antigen first described and defined by Kunin. We hope to show further that mutants deficient in enterobacte- rial common antigen (of the rfe- and the r f f type) lack the GlcN/ManNUA polymer (ECA- mutants) even in the PL fractions (unpublished results).
The amphipathic character of enterobacterial com- mon antigen makes it very surface-active. It readily attaches to lipopolysaccharide or cellular components, to erythrocytes and lymphocytes [4,48], thus providing the material with specificity for enterobacterial com- mon antigen. This might be a reason for the conflict- ing data on its isolation and localisation.
ManNUA is an essential constituent of entero- bacterial common antigen and modification of this constituent in the polymer by esterification or reduc- tion of the carboxylic group leads to a dramatic change of its serological behaviour (precipiting line in immunoelectrophoresis, hemagglutination inhibition capacity). This aminuronic acid is a rare constituent in gram-negative bacteria. It was described as consti- tuent of the K7 and the K56 antigen of E. coli [40,49]
and of the K15 antigen of Vibrio parahaemolyticus [SO]. In gram-positive bacteria ManNUA was de- scribed as part of the cell wall polysaccharide of M . lysodeikticus [42,51] and as a surface antigen of Staphylococcus aureus [52]. The biosynthesis of Man- NUA in E. coli was investigated by Kawamura [S3].
UDP-GlcNAc is first epimerized to UDP-ManNAc and then converted by a dehydrogenase to UDP- ManNAcUA, the activated form of ManNUA. This allows the labeling in vivo of both amino sugars (GlcN and ManNUA) by growing the bacteria in the presence of labeled GlcNAc as described previously by Kiss [21]. Labeled enterobacterial common antigen will certainly be of value for further studies on the biochemistry and the biology of this antigen.
We thank Mr R. Warth for analyses in the amino acid ana- lyzer, Miss H. Kochanowski for analytical ultracentrifugation, and
Mrs B. Straub and Mr D. Borowiak for skilful technical assistance.
We are indebted to Drs I . Fromme, C. Galanos, S. Hase, E. Th.
Rietschel and 0. Liideritz for experimental help and most valuable discussions, and to Dr S. Schlecht for mass cultivation of bacteria.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Makela, P. H. & Mayer, H. (1976) Bucteriol. Rev. 40,591 -632.
Le Minor, L., Chalon, A,-M. & Veron, M. (1972) Ann. Inst.
Whang, H. Y. & Neter, E. (1967) Proc. Soc. Exp. B i d . Med.
Kunin, C. M., Beard, M. V. & Halmagyi, N. E. (1962) Proc.
Kunin, C. M. (1963) J . Exp. Med. 118, 565-586.
Suzuki, T., Gorzynski, E. A. & Neter, E. (1964) J . Bucteriol.
Hammarstrom, S . , Carlsson, H. E., Perlmann, P. & Svensson, Johns, M. A., Whiteside, R. E., Baker, E. E. & McCabe, W. R.
McLaughlin, J. C. & Domingue, G. J. (1974) Immunol. Com- Marx, A. & Petcovici, M. (1975) Zentrulhl. Bukteriol. Puru- Neter, E. (1969) Curr. Top. Microhiol. Immunol. 47, 82- 124.
Mayer, H. & Schmidt, G . (1971) Zentrulbl. Bukteriol. Pura- sitenkd. Infektionskr. Hyg. Aht. I Orig. Reihe A , 216, 299- 313.
Mayer, H., Schmidt, G., Whang, H. Y. & Neter, E. (1972) Infect. Immun. 6, 540- 544.
Schmidt, G., Mannel, D., Mayer, H., Whang, H. Y. & Neter, E. (1976) J . Bucteriol. 126, 579-586.
Makela, P. H. & Mayer, H. (1974) J . Bacteriol. 119, 765-770.
Schmidt, G., Mayer, H. & Makela, P. H. (1976) J . Bucteriol.
Makela, P. H., Jahkola, M. & Luderitz, 0. (1970) J . Gen.
Eriksson-Grennberg, K. G., Nordstrom, K. & Englund, P.
Whang, H. Y., Mayer, H., Schmidt, G. & Netcr. E. (IY72) In- Schlecht, S. (1975) Zentrulbl. Buktrriol. Purtrsitenkd. Infok- Kiss. P. (1976) Thesis. University Freiburg i. Br.
Westphal, O., Liideritz, 0. & Bister, F. (1952) Z . Ntrtur/or..vc~/i.
Pasteur (Paris) 123,761 -774.
124,919-924.
Soc. Exp. Biol. Med. 111, 160-166.
98,1240-1243.
S. (1971) J . Exp. Med. 134, 565-576.
(1973) J . Bacteriol. 110, 781 -790.
mun. 3, 51 -75.
sitenkd. Infektionskr. Hyg. Aht. I Orig. 233,486-494.
127, 755- 762.
Microhiol. 60, 91 - 106.
(1971) J. Bucteriol. 108, 1210-1223.
,f&. Immun. 6, 533-539.
tionskr. H j ~ ~ y . Aht. I Orig. Reihe A . 232, 61 - 72.
7h, 148-155.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35 36 37 38 39 40 41
42 43 44 45 46.
47.
48.
49.
50.
51.
52.
53.
D. Mannel and H. Mayer, Max-Planck-Institut fur Immunohiologie, Stiibeweg 51, D-7800 Freihurg-Zahringen, Federal Republic of Germany
Galanos, C., Liideritz, 0. & Westphal, 0. (1969) Eur. J . Bio- Dankert, M., Wright, M. A,, Kelley, W. S. & Robbins, P. W.
Galanos, C. & Liideritz, 0. (1975) Eur. J . Biochem. 54, 603- Mayer, H. & Westphal, 0. (1968) J . Chromutogr. 33, 514- Hestrin, S. (1949) J . Biol. Chem. 180,249-261.
Rietschel, E. Th., Gottert, H., Liideritz, 0. & Westphal, 0.
Fromme, I. & Beilharz, H. (1978) Anal. Biochem. 84. 347-353.
Stellner, K., Saito, H. & Hakomori, S. I . (1973) Arch. Biochem.
Heath, E. C., Ghalambor, M. A. (1963) Biochem. Biophys. Res.
Lowry, 0. H., Roberts, N. R., Leiner, K. Y., Wu, M. L. &
Yphantis, D. A. (1960) Ann. N . Y . Acud. Sci. 88, 586-601.
Kratny, O., Leopold, H. & Stabinger, H. (1973) Methods En- Taylor, R. L. &Conrad, H. E. (1972) Biochemistry, 11, 1383- Neter, E. (1956) Bucteriol. Rev. 20, 166- 188.
Neter, E. (1965) Puthol. Microhiol. 28, 859-877.
Ouchterlony, 0. (1958) Prog. Allergy, 5, 1 - 77.
Scheidegger, J. J. (1955) Int. Arch. Allergy, 7, 103-110.
Mayer, H. (1969) Eur. J . Biochem. 8, 139-145.
Galanos, C., Liideritz, O., Rietschel, E. Th. & Westphal, 0.
(1977) in International Review of Biochemistry, Biochemistry of Lipids I I , 14, 239-335.
chem. 9,245 - 249.
(1966) Arch. Biochem. Biophys. 116, 425-435.
610.
525.
(1972) Eur. J. Biochem. 28, 166- 173.
Biophys. 155,464-472.
Commun. 10, 340-345.
Farr, A. L. (1954) J . Biol. Chem. 207, 10-17.
zymol. 27, 98 - 110.
1388.
Perkins, H. R. (1963) Biochem. J . 86,475-483.
Brown, D. H. (1957) Methods Enzymol. 3, 158.
Erbing, Ch., Granath, K., Kenne, L. & Lindberg, B. (1976) Nasir-Ud-Din & Jeanloz, R. W. (1976) Curbohydr. Res. 47, BeMiller, J. N. & Kumari, G. V. (1972) Curbohydr. Res. 25, Kiss, J. (1974) Curbohydr. Res. 29, 229-303.
Whang,H. Y., Cohen,E. & Neter,E. (1965) Vox Sung. 10, Flemming, H. C. (1972) Diploma Work, University of Freihurg.
Torii, M., Sakakibara, K. &Kuroda, K. (1973) Eur. J . Biochem.
Hase, S. & Matsushima, Y. (1970) J . Biochem. ( T o k y o ) 68, Wu, T. C. M. & Park, J. T. (1971) J . Bucteriol. 108,874-884.
Kawamura, T., Ichihara, N., Ishimoto, N. & Ito, E. (1975) Curhohydr. Res. 47, 5 - 7.
245 - 260.
419-429.
161 - 168.
37,401 -405.
723 - 730.
Biochem. Biophys. Res. Commun. 66, 1506- 1512.
